Vers une meilleure modélisation de l'évapotranspiration (ET) sous des conditions environnementales diversifiées

Authors: Hajji, Islem
Advisor: Anctil, François; Music, Biljana
Abstract: The process of evapotranspiration (ET) is one of the most significant components of the earth water cycle. It is the most important consumer of energy, creating the link between the cycles of water and energy of the Earth. Therefore, control of this process is essential to better understand the complex mechanisms governing the interactions between the soil and the atmosphere in terms of mass and heat transfer. Given its major interest in various disciplines,the study of such a process has been the subject of several scientific works in recent years.This has led to the development of a wide variety of methods for quantification of ET and its main components, such as evaporation and transpiration. Nevertheless, despite the scientific effort, precise ET estimation remains limited by a number of unresolved problems. Some of these problems are related to the variability of ET in space and time, as well as in climate sand biomes. Other problems are related to the setting and calibration of the proposed ET models and the important number of data needed to run them.To overcome difficulties related to existing ET approaches, a new model, called the Maximum Entropy Production (MEP) model, was recently developed based on out-of-equilibrium thermodynamics and Bayesian probability theory. This model offers an interesting alternative to efficiently estimate ET and the surface heat fluxes, using fewer input variables than the existing models, while respecting the closure of the surface energy balance. There are three MEP versions used to separately estimate evaporation, transpiration and sublimation. However, until now, the capabilities of this model have only been evaluated in relatively simple and homogeneous sites and climatic conditions. Its applicability on complex sites with varied climates and plant covers has not been studied. In this context, the present project aims to investigate the capabilities of the MEP model to accurately predict ET in a wide range of climates and plant covers and to propose improvements in order to enhance such capabilities if necessary. To achieve its objectives, this work is divided into two parts.The first part of the work has focused on the investigation of the MEP model under complex climatic and plant conditions in the absence of snow effects. To this end, a coupling between the two versions of the MEP model used to estimate soil evaporation and vegetation transpiration has been developed. The obtained coupling model, which is referenced to in the followingas MEP-ET model, is based on the introduction of a weighting coefficient characterizing the proportion of the vegetated surface with respect to the total surface of the studied site. Using this model, total ET in partially vegetated sites is calculated as the weighted sum of soil evaporation and vegetation transpiration. To study the performance of the MEP-ET model,it has been applied to eight FLUXNET sites, with different climatic and plant conditions, in the American continent. The comparison between the obtained estimates and the in-situ observations of the ET shows the great ability of the MEP-ET model to estimate the ET in sites characterized by an abundant to moderately limited presence of water. In sites characterizedby conditions of high water stress, the estimates obtained are far from being realistic. This is because the original version of the MEP-ET model does not take into account the effects of soil moisture stress on stomata. To overcome this problem, a new regulation parameter has been introduced to correctly reproduce these effects. The addition of this parameter has significantly minimized biases of the original version of the MEP-ET model, especially underwater stress conditions. The generalized MEP-ET model, involving the regulation parameter,is much more efficient than other models conventionally used to estimate ET at all test sites, such as the Penman–Monteith (PM), modified Priestley–Taylor–Jet Propulsion Laboratory(PT-JPL), and air-relative-humidity-based two-source model (ARTS).The second part of this work has focused on the inclusion of snow effects on the ET estimationin order to allow for simulation of the surface water vapor fluxes (latent heat) over the entire snowpack cycle, including snow accumulation and snow melting at the beginning of the growing season. To do this, the generalized MEP-ET model proposed in the first part of the work has been coupled with the MEP version for snow to consider effects of water loss in the atmosphere during the snowpack life cycle on the ET. Two hypotheses based on multi-year observations of several sites have been used to develop this coupling: (1) effects of sublimation are negligible during melting when the snowpack is isothermal (0°C) and (2) effects of transpiration are progressively activated according to the air temperature when the vegetation awakes. The proposed coupling approach is shown to be very effective in modeling the total surface watervapor fluxes during the snowpack life cycle.
Document Type: Thèse de doctorat
Issue Date: 2020
Open Access Date: 21 December 2020
Grantor: Université Laval
Collection:Thèses et mémoires

Files in this item:
Description SizeFormat 
36065.pdf3.66 MBAdobe PDFThumbnail
All documents in CorpusUL are protected by Copyright Act of Canada.